The benefits of rigid implants (i.e., internal fixation) in the axial skeleton include rigid stabilization, maintenance of alignment, minimal postoperative immobilization, earlier return to function, and potentially enhanced fusion rates. A potential shortcoming of rigid implants is that they may stress-shield the bone graft and result in nonunion, implant failure, or both. Stress shielding refers to an implant-induced reduction of bone healing–enhancing stresses and loads, resulting in stress reduction osteoporosis or nonunion ( Fig. 63-1 ). This hypothesis is in keeping with Wolff’s law, which postulates that the form and function of bone is a result of changes in the internal architecture according to “self-ordered” mathematical rules. In contemporary terms, skeletal morphology is substantially controlled by mechanical function, and bone remodeling, both locally and throughout the skeleton, is influenced by the level and distribution of the functional strains within the bone. A corollary to Wolff’s law is that bone heals optimally under compressive, as opposed to tensile, forces. Experimental studies in the thoracolumbar spine show that a 70% or greater axial load should be transmitted through the spine, not the implant, optimally to enhance arthrodesis and provide acute stability.

A ventral rigid cervical implant caused stress shielding. This resulted in nonunion (pseudarthrosis) in a patient with preexisting osteoporosis, as depicted. Arrows indicate the location of the nonunion (pseudarthrosis). A, Lateral radiograph; B, close-up.

(From Benzel EC: Biomechanics of spine stabilization, Rolling Meadows, IL, 2001, American Association of Neurological Surgeons.)

In an attempt to improve on the shortcomings of rigid implants, there has been a resurgence of interest in dynamic implants, particularly for use in the cervical spine. The concept of dynamic implants is not new. Dynamic hip arthroplasties have been employed successfully for femoral neck fractures. These dynamic implants allow for the femoral neck to shorten or collapse along its axis so that the bone is subject to optimal bone-healing compressive forces. Advocates of dynamic implants hypothesize that implants that permit a limited and controlled type of deformation may be desirable. Some experts have termed this controlled dynamism . In the spine, allowing for some axial deformation but not angular deformation (kyphosis) may be optimal. Occasionally, the failure of a rigid implant may permit fusion because the bone graft and vertebral bodies are subsequently exposed to the appropriate bone healing–enhancing forces. In this case, the implant has “dynamized by failing” ( Fig. 63-2 ).

Failure of an implant (by fracture) may allow fusion to occur. In a sense this implant, dynamized by failure, as depicted, allows the bone graft to see bone healing–enhancing compression forces.

(From Benzel EC: Biomechanics of spine stabilization, Rolling Meadows, IL, 2001, American Association of Neurological Surgeons.)

The first ventral cervical plate and screw system was introduced by Bohler in 1964. This system ultimately culminated in the development of the Caspar (Aesculap, Center Valley, PA) and Orozco (Synthes, West Chester, PA) plate systems in the early 1980s. These early ventral cervical plates were dynamic implants and are classified as having unrestricted backout properties (i.e., nonlocking and nonrigid) because of a lack of fixation at the screw-plate interface. These implants permit a significant transfer of load through the bone graft, increasing the likelihood of fusion. The nonfixed moment arm nature of the screw caused degradation of the screw-bone interface with cyclic loading. This effect can be minimized with bicortical screw purchase, which requires C-arm fluoroscopy. The main disadvantage of these plates is that the nonlocking and nonrigid (i.e., variable angle) screws led to high rates of screw backout and screw breakage with graft subsidence ( Fig. 63-3 ).

Screws may fracture as a result of excessive stresses placed on them by the subsiding spine, as depicted. A, Anteroposterior radiograph. B, Lateral radiograph. The screws positioned in holes fractured; they could not axially subside. The screws positioned in slots maintained fixation while dynamizing, permitting, and encouraging fusion.

(From Benzel EC: Biomechanics of spine stabilization, Rolling Meadows, IL, 2001, American Association of Neurological Surgeons.)

The next generation of ventral cervical plates included the CSLP (DePuy Synthes, Raynham, MA) and Orion (Medtronic Memphis, TN). The CSLP was developed by Morscher in Europe in the early 1980s and introduced in the United States in the early 1990s. The major advantage of this generation of devices is that they do not require bicortical screw purchase. The CSLP uses a titanium expansion screw that rigidly secures the screw to the plate, greatly reducing the incidence of screw backout. In contrast to the Caspar plate where screw angulation could be varied, the CSLP has a predetermined (rigid) screw trajectory, which is perpendicular at the caudal end and 12 degrees rostrally. It has been suggested that these types of restricted, constrained plate-screw configurations are preferable in trauma cases, in which immediate stability is desired; however, this concept remains unproved.

One concern with rigid plates such as the CSLP and Orion is that they were thought to stress-shield the bone graft by reducing the compressive forces that the bone graft experiences and result in increased rates of nonunion (pseudarthrosis). This concern led to interest in the design of dynamic implants. These newer-generation dynamic implants improved on the Caspar plate design by preventing screw backout while allowing for some movement at the plate-screw interface. This dynamism allowed for compressive forces to be shared between the implant and the bone graft—so-called load sharing. Dynamic cervical plates can be classified into rotational or translational, depending on the type of movement that is permitted at the plate-screw interface. The translational dynamic plates also can be subdivided further into internally and externally dynamized plates.